Free SAA-C03 practice scenarios: SQS decoupling

Buffering work so nothing is lost — FIFO ordering, visibility timeout, dead-letter queues and scaling on backlog. Six scenarios, SAA-C03, 2026 edition.

Try 6 free SQS decoupling scenarios in the exam's own style — every answer explained, no sign-up.

Straight to the 6 free questions ↓

SQS is the exam’s default answer for decoupling — putting a durable buffer between a producer and a consumer so a burst, a slow consumer, or an offline consumer never loses work. The scenarios then test the queue’s specific behaviours.

Standard vs FIFO

The first fork is the queue type. A standard queue offers near-unlimited throughput with best-effort ordering and at-least-once delivery (duplicates possible). A FIFO queue guarantees strict ordering and exactly-once processing. When a stem says “exact order” or “no duplicates” — payments, sequential events — it is FIFO. Otherwise standard.

Visibility timeout and dead-letter queues

Two mechanisms manage in-flight and failing messages:

Visibility timeout — when a consumer receives a message, SQS hides it from other consumers while it is processed. Delete on success; if not deleted (a failure), it reappears for retry. Set it longer than your processing time.
Dead-letter queue (DLQ) — messages that exceed a maximum receive count are moved aside, so poison messages stop clogging the queue and can be inspected without blocking healthy work.

Scaling on the backlog

To make the consumer fleet track the workload, scale the Auto Scaling group on the queue depth (the number of messages waiting) via a CloudWatch metric — more workers when the backlog grows, fewer when it drains. A fixed fleet sized for peak wastes money; manual scaling defeats the purpose.

Long polling

When a queue is polled constantly but often empty, long polling makes each receive wait for a message to arrive rather than returning immediately empty — cutting empty responses and cost. Short polling does the opposite and is the distractor here.

How to use the six scenarios below

Answer each one before opening the explanation, and when you miss, name the behaviour the requirement needed — “exact order”, “in-flight exclusivity”, “isolate failures”, “track the backlog”, “fewer empty responses”. Every explanation states that behaviour and dismantles the distractors, which are usually another SQS feature solving a different need. For the domain overview, see Design Resilient Architectures.

An Amazon SQS queue decoupling a producer from a consumer. Messages sit durably in the queue until a consumer polls them; a visibility timeout hides an in-flight message from other consumers while it is processed, and a message deleted only after success. A dead-letter queue captures messages that repeatedly fail. A standard queue offers best-effort ordering and at-least-once delivery at near-unlimited throughput, while a FIFO queue guarantees exact ordering and exactly-once processing. An Auto Scaling group scales the consumer fleet on the queue's backlog depth.

6 free SAA-C03 practice questions

Answers and explanations — no email wall

SAA-C03 Question 1 of 6

A payments pipeline requires that messages are processed in the exact order they were sent and that each message is processed exactly once, with no duplicates. Which Amazon SQS queue type meets this?

Answer: A — A FIFO queue.

A FIFO queue guarantees strict ordering and exactly-once processing — precisely the two requirements for a payments pipeline. Why not the others: a standard queue provides only best-effort ordering and at-least-once delivery, so duplicates and out-of-order messages are possible — a visibility timeout does not change that. An SNS topic is publish/subscribe for fan-out, not an ordered queue. When the stem says 'exact order' and 'exactly once', it is a FIFO queue.

SAA-C03 Question 2 of 6

A consumer retrieves a message from an SQS queue and takes some time to process it. During that time, the message must not be delivered to any other consumer. Which SQS mechanism provides this?

Answer: A — The visibility timeout.

When a consumer receives a message, SQS hides it from other consumers for the duration of the visibility timeout; if the consumer finishes and deletes it in time, it is gone, otherwise it reappears for retry. That is exactly 'not delivered to another consumer while processing'. Why not the others: the retention period is how long an unconsumed message stays in the queue before expiring. Long polling reduces empty receive responses. A dead-letter queue captures repeatedly failing messages. In-flight exclusivity is the visibility timeout.

SAA-C03 Question 3 of 6

Some messages in an SQS queue fail processing repeatedly and are clogging the queue and consuming retries. The team wants these problem messages set aside for later inspection without blocking healthy processing. What should they configure?

Answer: A — A dead-letter queue with a maximum receive count.

A dead-letter queue receives messages that exceed a configured maximum receive count, moving poison messages off the main queue for inspection while healthy messages flow — exactly the requirement. Why not the others: a longer visibility timeout just delays each retry, it does not isolate the failing messages. Switching to FIFO changes ordering semantics, not failure handling. A shorter retention period would silently drop messages, losing the ones you want to inspect. Isolating repeatedly failing messages is a dead-letter queue.

SAA-C03 Question 4 of 6

A fleet of EC2 consumers processes an SQS queue whose backlog grows and shrinks sharply during the day. The company wants the number of consumers to track the amount of work waiting. What is the best approach?

Answer: A — Scale the consumer Auto Scaling group on the queue depth (number of messages) via a CloudWatch metric.

Scaling the Auto Scaling group on the SQS queue depth (ApproximateNumberOfMessages) through a CloudWatch metric makes the consumer count follow the backlog — more workers when work piles up, fewer when it drains. Why not the others: a fixed fleet sized for peak wastes money most of the day and cannot flex. Manual scaling is exactly the toil automation removes. (Queue depth is the standard signal here; the point is to scale on the backlog automatically, not by hand.) Backlog-driven scaling uses the queue depth metric.

SAA-C03 Question 5 of 6

An SQS queue is polled continuously by consumers but is often empty, generating many empty responses and unnecessary API cost. How can this be reduced?

Answer: A — Enable long polling so the consumer waits for messages to arrive before returning.

Long polling makes a receive request wait (up to a set time) for a message to arrive instead of returning immediately empty, which cuts the number of empty responses and the associated cost and API calls. Why not the others: short polling returns immediately and produces exactly the empty responses the requirement wants to reduce. The visibility timeout governs in-flight messages, not polling behaviour. A FIFO queue changes delivery semantics, not empty-receive cost. Reducing empty responses is long polling.

SAA-C03 Question 6 of 6

An image-upload service must hand off processing to a downstream service that is periodically taken offline for maintenance. No uploads may be lost, and the downstream service should resume processing when it returns. What should sit between them?

Answer: A — An SQS queue that durably holds messages until the downstream service consumes them.

An SQS queue durably stores messages while the downstream service is offline and lets it consume them when it returns, so nothing is lost during maintenance — decoupling producer from consumer. Why not the others: a direct synchronous call fails whenever the downstream service is down, losing the work. A load balancer distributes requests to available targets but does not store work for an offline service. An in-memory cache is not durable and would lose messages on restart. Durable hand-off across an offline consumer is an SQS queue.

That is exactly how every question in the course works — answer, explanation, why-not. The real set continues in the practice player: 10 free questions, no sign-up.

Continue with the 10 free questions →

Those 6 questions were the start.

The exam does not test whether you recognise a term — it tests whether you can rule out three plausible answers under time pressure. That is what the explanations above are for, and there are 780 more questions built exactly like them.

Collecting questions yourself

Scattered across forums, of unknown age
Answer keys without reasoning
No idea which domain you are weak in

Practising with a system

780 questions in 12 full tests, SAA-C03 (2026)
Every option explained — including the wrong ones
Readiness per exam domain, and drills for your weakest

Start free — 10 questions

Straight into the player. No account, no email.

Frequently asked questions

Standard or FIFO queue — how do I choose?

Use a standard queue for maximum throughput when best-effort ordering and at-least-once delivery are acceptable. Use a FIFO queue when the requirement is strict ordering and exactly-once processing — payments and other order-sensitive pipelines. The words 'exact order' or 'no duplicates' point to FIFO.

What is the visibility timeout for?

It hides a message from other consumers while one consumer processes it. Set it longer than your processing time so a message is not picked up twice; if processing fails and the message is not deleted, it reappears for retry after the timeout.

Are these real exam questions?

No. They are original scenarios written for this page in the exam's style and difficulty. Reproducing real exam items violates the AWS certification agreement and can cost a candidate their certification — ours come with the reasoning attached instead.

What is the one-line takeaway?

SQS buffers work durably so nothing is lost; FIFO for exact order and exactly-once, visibility timeout for in-flight exclusivity, a dead-letter queue for poison messages, and scale consumers on queue depth.

Updated for SAA-C03 (July 2026). The sample questions above are our own work in the style of the exam — not real exam items. The exam itself is set and marked by the certification body.